Sawyer, R. J., Graham, S. & Harris, K.R. (1992). Direct teaching, strategy instruction, and strategy instruction with explicit self-regulation: Effects on the composition skills and self-efficacy of students with learning disabilities. Journal of Educational Psychology, 84(3), 340-52.
Chosen Study
This study is noteworthy because of its contribution to the field of self-regulated learning studies with students with learning disabilities (LD). The basis of this study is anchored on improving academic skills using cognitive strategy instruction. The study’s simple but straightforward methodology merits review by novice researchers. The statistical analysis was simple and easy to understand. This research is a good model to review when searching for examples of how to show the effectiveness of multicomponent strategy intervention models for students with LD.
Problem
Sawyer, Graham & Harris sought to address differential effects in the writing of students with LD in four experimental conditions: (1) full self-regulated strategy development (SRSD); (2) SRSD without explicit self-regulation, (3) direct teaching of strategy usage, and (4) nonrandomized practice-control. In this study, direct teaching was defined as teaching without (1) explicit self-regulation procedures, (2) the use of personalized self-instructions to guide the use of strategies, (3) modeling by teacher, and (4) collaborative practice with the teacher.
This study replicates and extends a previous study conducted by Graham and Harris (1989b). The results showed that both strategy instruction conditions resulted in meaningful and lasting effects on the schematic structure and quality of students’ compositions. Building from the Graham and Harris (1989b) study, the authors of this study sought to determine whether explicit self-regulation meaningfully augments strategy instruction when combined with instruction in the significance of the strategy.
The study design enabled the identification of the relative contributions of instructional components responsible for the change in learning. Separating the components allowed the authors to measure the differential effects of the instructional types at posttest, generalization and maintenance. New to this study was the addition of a practice control group. The authors also used a social validation condition of normally achieving students for comparison. The authors were able to see differential effects in schematic structure and quality and levels of self-efficacy in students with LD when compared to the control group and normally achieving students in varying conditions.
Methods
Three groups of students were selected for the study. Thirty-three fifth- and sixth grade students with LD were randomly assigned to the three strategy instructional conditions. Another 10 randomly selected fifth- and sixth-grade students with LD participated in a nonrandomized practice-control condition.
Students with LD were identified according to the following criteria: (a) identification as LD by their local school district, (b) IQ scores between 80 and 120 on a norm-referenced intelligence test, (c) achievement that was at least 2 years below grade level in one or more academic areas, (d) absence of other handicapping conditions, and (e) interviews with the resource teacher indicating significant composition problems were evident. Data reported a mean grade equivalent delay of approximately 2 years in both reading and writing. For sixth grade students, reading delay was 2.6 years with a writing delay of 2 years.
Graham and Harris used several instruments to measure the different effects of the study. A story grammar scale was used to measure sensitive effects of an instructional program. This scale was deemed valid in several separate investigations. The scale was shown to be highly correlated with other measures assessing the overall quality in students’ written work. To assess the quality of students’ stories, a holistic rating scale using a 7-point scale was created. Samples of high-, medium- and low-scoring compositions were used in grading the students’ writing sample. Two former elementary-school teachers who were unfamiliar with the study scored the writing samples. They were trained in using the holistic rating scale. The 2 scores were averaged and reported.
The authors used measures developed by Graham and Harris (1989b) for assessing students’ self-efficacy in writing. The instrument was designed in accordance with procedures outlined by Bandura and Schunk (1981). The participants were oriented to the scale using age appropriate examples and were asked to use the scale to rate their self-efficacy. Students were asked 10 self-efficacy related questions. Scores were averaged and reported.
To assess the usage of strategies taught in the instructional conditions, notes and the written stories were examined to obtain direct evidence.
To obtain the writing samples, authors used black-and-white pictures for the writing stimuli during five different writing probes. These probes were given for pre-test and posttest; generalization probe given in the student’s classroom after the end of instruction; and two maintenance probes given at 2 and 4 weeks after instruction. These pictures were validated at a preplanned pilot study. The authors were diligent at establishing rigor with their instruments and put in place a number of procedures to ensure reliable and valid outcomes. This includes ensuring that there is no cumulative effect by rotating pictures at each writing probe.
Instruction was given in the three strategy instruction conditions by four preservice teachers with extensive field experience and who were majoring in education. The teachers were not informed of specific purpose of the study to hedge against predisposed affinity to a particular strategy. Instruction was given in the 4 small groups in each condition. Sessions were three instructional sessions per week for approximately 3 weeks. At pretest and posttest, self efficacy measure and writing probes were each administered on different days to avoid carryover effects. Maintenance probes were administered to students with LD who received the three strategy instruction conditions. Generalization probes were administered only to the students with LD in the strategy instruction conditions within 1 week of the end of the strategy instruction. The normally achieving students were administered only one writing probe and collected when the students with LD in strategy instruction completed their posttest.
Students in only the full SRSD condition received instruction and practice designed to facilitate goal setting and self-monitoring in addition to the other SRSD strategies. They also received full feedback. Students in the SRSD condition without explicit self-regulation instruction received all the strategies except for goal setting and self-monitoring. There were no conferences given to receive feedback from the instructor. In direct teaching students did not develop practice using self-instructions to help them regulate the use of strategies. There was no overt modeling and collaborative work among the students. There was no explicit instruction in goal setting and self-monitoring.
For the non-randomized control group, writing instruction was not provided. Participants were asked to write three stories independently and given no assistance in writing. Feedback was also withheld.
Results
Means and standard deviations on story grammar scale were reported for all four conditions, practice control and norm group. In all writing probe cases, the mean for the full SRSD condition increased compared to direct teaching method and SRSD without explicit goal setting and self-monitoring instruction. The posttest means for the full SRSD condition were also higher when compared to the normal achieving group and practice control group.
Means and standard deviations on the writing performance also show that full SRSD is an effective instructional strategy to improve and generalize writing skills. For all writing probes, means for the full SRSD condition increased when compared to the other conditions. The posttest means for the full SRSD condition were also higher when compared to the normal achieving group and practice control group.
Evidence of strategy use showed 55% of students with LD in the direct teaching condition, 64% in the SRSD condition without explicit self-regulation and 73% of the full SRSD showed overt evidence of using the strategy when writing stories. Self-efficacy pre and posttest showed no significant difference.
Discussion
The research investigated component analysis, writing performance, and self-efficacy. Students who received full SRSD and SRSD without explicit strategy instruction conditions received significantly higher schematic structure than did students in the practice-control condition. No augmentive effects were found on schematic structure for goal setting and self-monitoring. However, the results highlighted the use of self-statements, modeling and collaborative practice was beneficial to students with LD. Generalization and maintenance results showed that the full SRSD instruction benefited students with LD. The writing performance showed improvement with the full SRSD instruction. Scores among students with LD from all three instructional groups were not significantly different from the normally achieving students. In contrast, the normally achieving students did have significantly higher scores than students with LD in the practice-control condition. This result is important given the typical extreme differences in writing scores of students with LD and the normally achieving peers. This shows that SRSD instruction in writing is beneficial for students with LD. Finally, measures of self-efficacy showed no significant differences between pre and posttest measures.
Critique
A key strength of this study is the decomposition of strategy instruction to enable the identification of the specific variable that contributed to increased learning. Data were gathered to emphasize the effect of each particular component. In addition, the clever use data from the nonrandomized practice-control group and the normally-achieving group provided a nice comparison group to stress the benefits of self-regulated strategy instruction. The lack of a sensitive measurement tool to accurately assess self-efficacy of students with LD is a relative weakness of the study. The self-efficacy results merely confirmed what previous literature stated regarding the mismatch of self-efficacy and achievement by students with LD. The authors cited many opportunities for study in the area of self-efficacy. Unfortunately, the study was not able to add to the body of self-efficacy literature.
Implications
The major implication of this study is that explicit self-regulated strategy instruction with goal setting and self-monitoring is a viable strategy to improve story grammar and quality of writing among fifth- and sixth-grade students with LD. Results of this study showed that full SRSD instruction met all three components of effective strategy instruction: (1) targets strategies; (2) includes metastrategy information; and (3) develops self-regulation skills needed to use, monitor, maintain and generalize strategies effectively.
The study was designed to compare the results of four treatment conditions against results of the practice-control group and the normally achieving validation group. Additionally, data were gathered at five different points in time to assess generalizability and sustainability of skills. This design mapped out the impact of full SRSD instruction with writing. Future research could use this method when assessing the impact of full SRSD instruction in other academic content areas such as reading comprehension and math.
Practical implications of this study points to the viability of SRSD instruction in the classroom. Evidenced by improved measures in grammar and writing performance, self-regulated strategy instruction with goal setting and self-monitoring is cost effective in terms of time and materials. Students’ report on the use of strategies shows that the strategy is effective for sustaining desired writing outcomes. For these reasons, practitioners will be more apt to consider including these strategies in their teaching practice.
